Nowadays, so-called “crowd computing” technologies are provided to allow people to use various kinds of services via servers and storage systems on some network. According to such crowd computing technologies, as a lot of users save various kinds of data on that storage system on the network, the amount of data accumulated there should keep on skyrocketing from now on. Meanwhile, to preserve such a huge amount of data saved, it should also be increasingly important to devise a method for saving that enormous amount of data as securely and as reliably as possible. And as such a huge amount of data should now be saved on the network with as much reliability as possible, an optical read/write apparatus to be used as such a storage system is required to verify the data that the apparatus is writing.
An optical read/write apparatus which performs a write operation and a verify operation using one, two or more optical heads (or optical pickups) is disclosed in Patent Document No. 1, for example. A technique for performing a write operation and a read operation for verification purposes in parallel is called a “DRAW (direct read after write)” technique.
According to the conventional verification method, when data that has been written on a storage medium is read in a predetermined unit, the decision is made whether the frequency of errors that have occurred in the data yet to be subjected to “error correction” is equal to or greater than a predetermined reference value or not. If a mark that has been recorded on a storage medium has an inappropriate shape, a read error will occur. If the frequency of occurrence of such read errors is sufficiently low, the data represented by those recorded marks can still be restored by error correction and can be read as intended. However, if the frequency of occurrence of such read errors is too high, the data cannot be restored and read properly even when subjected to the error correction. Thus, if the frequency of occurrence of read errors during a verify operation is equal to or greater than a reference value, the decision is made that that data should be rewritten.